The Human Genome Project took several years to complete, yet it is only the beginning of a period in which large amounts of DNA and RNA sequence information will be required for medical diagnostics, forensics, and developmental biology. Conventional sequencing technology has limitations in cost, speed, and sensitivity and the demand for sequence information far outstrips the current ability to obtain it. We recently demonstrated the first proof of principle experiments for a new technology that will provide a fast, low cost, and highly parallel technique for DNA and RNA sequencing. This technology uses single molecule detection of fluorescently labelled nucleotides after DNA polymerase incorporates labeled dNTPs into immobilized individual DNA molecules. A major advantage of this technique over current sequencing methods lies in its ability to obtain sequence information from millions of independent molecules in parallel. Here we propose to develop reagents and methods for single molecule sequencing runs with longer read lengths and higher accuracy, ultimately reaching the NIH gold standard of 99.99%, while reducing the cost of sequencing a mammalian genome to below $100,000. [unreadable] [unreadable]